This is perhaps a little naive, but without aircraft flying into a hurricane, how can you determine wind speed to a plus or minus 5mph accuracy as is being reported with Bertha? I realize the shape of the eye, cloud top temperature,etc can be used to determine relative strength, but can these wind speed numbers truly be trusted without onsite evaluation?
Posted July 14, 2008
MIKE MOSS SAYS: Nothing naive at all about your question and observation, Kenneth. You are right on target that observations via satellite of storm structure, organization, cloud-top temperature, temporal trends and more are often the only information that Tropical Cyclone forecast centers have to go on in assigning an intensity to a storm, and that there are certainly sources of ambiguity that can result in errors in estimating maximum surface wind. The process is most challenging when storms are rapidly intensifying or weakening, are undergoing eyewall replacement cycles, have a cirrus overcast obscuring the center of rotation, or are located where the satellite is viewing the storm at a significant slant angle.
Having said all that, a great deal of effort has gone into establishing increasingly objective methods of estimating the intensity since the early 1970s, and a technique was initially developed back then by meteorologist Vernon Dvorak. His methodology, with gradual modification through the years, has proven extremely valuable and surprisingly robust, and is still used today when aircraft reconnaissance and/or nearby ship reports are not available (true for most storms east of 55 deg W longitude) or in between recon sorties. The most recent major validation effort in which results from the technique were compared against concurrent on-site data resulted in the Dvorak estimates being within 5 knots of the measured maximum sustained wind speed roughly 50% of the time, 12 knots about 75% of the time and 18 knots about 90% of the time, and new observation methods such as microwave imaging, paired with improved computational capacities, promise to improve on those numbers even more. For a very nice overview of the technique, should you be interested in more detail, see